加载中…
加载中…
microsoft/codebert-base fine-tuned on the PECAN benchmark to identify the programming language of a raw 512-byte code fragment (319 languages, no file name or extension). Test split: accuracy 0.9575, macro-F1 0.9304. Artifact and scoring code: https://anonymous.4open.science/r/pecan-benchmark
microsoft/codebert-base